• 検索結果がありません。

A REMARK ON ASSOCIATED PRIMES IN THE COHOMOLOGY ALGEBRA OF A FINITE GROUP (Cohomology Theory of Finite Groups and Related Topics)

N/A
N/A
Protected

Academic year: 2021

シェア "A REMARK ON ASSOCIATED PRIMES IN THE COHOMOLOGY ALGEBRA OF A FINITE GROUP (Cohomology Theory of Finite Groups and Related Topics)"

Copied!
5
0
0

読み込み中.... (全文を見る)

全文

(1)

A REMARK ON ASSOCIATED PRIMES

IN THE COHOMOLOGY ALGEBRA OF A FINITE GROUP

TETSURO OKUYAMA

奥山 哲郎

HOKKAIDO UNIVERSITY OF EDUCATION, ASAHIKAWA CAMPUS

北海道教育大学・旭川校

Introduction

Let $G$ be a finite group and $H^{*}(G, k)$ be the cohomology algebra of $G$

over an

al-gebraically closed field $k$ of characteristic $p>0$. It is well known that $H^{*}(G, k)$ is

a

commutative noetherian graded k-algeba. In my talk we discussed properties of the

as-sociated primes in $H^{*}(G, k)$ and gave a slight generalizaton of

a

result of Green [7].

Theorem 0.1.

If

$\mathfrak{p}$ is

an

associated prime in $H^{*}(G, k)$ with $\dim H^{*}(G, k)/\mathfrak{p}=s$, then

there exists an elementary abelian p-subgroup $E$

of

rank $s$

of

$G$ such that the depth

of

$H^{*}(C_{G}(E), k)$ is $s$ and $\mathfrak{p}={\rm Res}_{G,E}^{-1}(\sqrt{0})$.

Conversely, let $E$ be an elementary abelian p-subgroup $E$

of

rank $s$

of

$G$ and assume

that the depth

of

$H^{*}(C_{G}(E), k)$ is $s$. Then ${\rm Res}_{G,E}^{-1}(\sqrt{0})$ is

an

associatedprime in $H^{*}(G, k)$.

The first assertion wasknown to be true using the notion of the SteenrodAlgebra [4].We

shall give a different proof using arguments by Carlson [3]. And the second assertion is

known to be true if $C_{G}(E)=G$ by Green [7].

1. The Rank-Restriction Condition

Let $r$ be thep-rank of $G$. Let $A_{s}=\mathcal{A}_{s}(G)$ be the set of elementary abelian p-subgroups

of $G(1\leqq s\leqq r)$ and set $\mathcal{H}_{s}=\mathcal{H}_{s}(G)=\{C_{G}(E) ; E\in A_{s} \}$. And for $1\leqq s\leqq r$,

let $\mathcal{K}_{s}=\mathcal{K}_{s}(G)$ be the family of subgroups $K$ of $G$ such that the Sylow p-subgroups of $C_{G}(K)$

are

not conjugate to

a

subgroup ofanyof the groups in $\mathcal{H}_{s}$. (Ifsuch

a

subgroup $K$

does not exists, then we set $\mathcal{K}_{s}=\{1\}.)$ Notice that any elementary abelian p-subgroup

of $K\in \mathcal{K}_{s}$ have rank at most $s-1$ . For

a

nonempty family $\mathcal{P}$ of subgroups of $G$, set

${\rm Im} Tr_{P,G}=\sum_{P\in P}{\rm Im}$ Tr$P,G$ and KerRes$G, \mathcal{P}=\bigcap_{P\in P}$KerRes$G,P$

.

Then by

a

theorem of

Benson (Theoreml.1 [1]) $\sqrt{{\rm Im} Tk_{\mathcal{H}_{s}G}}=\sqrt{Ker{\rm Res}_{G\mathcal{K}_{s}}}$.

Let $\zeta_{1},$

$\cdots,$$\zeta_{r}$ be a homogeneous system of parameters in $H^{*}(G, k)$ satisfying the

rank-restriction condition. Then for $1\leqq s\leqq r,$ ${\rm Res}_{G,E}(\zeta_{S})=0$ for any $E\in \mathcal{A}_{s-1}$,

we

see

that $\zeta_{s}\in\sqrt{Ker{\rm Res}_{G\mathcal{K}_{s}}}=\sqrt{{\rm Im} T\}_{?i_{s}G}}$. Thus replacing $\zeta_{s}$’s by a suitable p-power, we may

assume

that $(_{s}\in{\rm Im} Tk_{\mathcal{H}_{s},G}$ for each $1\leqq s\leqq r$

.

Thus we always have a homogeneous

sys-tem of parameters (1,$\cdots,$ $(_{r}$ in $H^{*}(G, k)$ satisfying the following conditions. See Lemma

8.5 [2] for the condition 3 and Lemma 3.4 and Theorem 9.6 [2] for the condition 4.

Condition 1.1.

(1) $\zeta_{1},$

(2)

(2) $\zeta_{s}\in{\rm Im} Tr_{\mathcal{H}_{s},G}$

(3) For each $s$ and $E\in \mathcal{A}_{s}$, the restrictions

of

$\zeta_{1},$

$\cdots,$ $\zeta_{s}$ to $H^{*}(E, k)$

form

a

homo-geneous system

of

parameters (and $(_{s+1},$ $\cdots,$ $(_{r}$ restrict to $0$). In particular, their

restrictions to $H^{*}(C_{G}(E), k)$

form

a regular sequence

for

$H^{*}(C_{G}(E), k)$.

(4)

If

$H^{*}(G, k)$ has depth $s$, then $\zeta_{1},$ $\cdots$ , $\zeta_{s}$

form

a regular sequence.

(5) $\dim H^{*}(G, k)/((s+1, \cdots, \zeta_{r})$ is $s$.

Lemma 1.2 (Carlson [3]). Let $0\neq\eta\in H^{*}(G, k)$ be a homogeneous element.

Assume $\dim H^{*}(G, k)/Ann_{H^{*}(G,k)}(\eta)=s$. Then the following statements hold.

(1)

If

$t>0$, then

for

any $H\in \mathcal{H}_{t},$ ${\rm Res}_{G,H}(\eta)=0$.

(2) For

some

$H\in \mathcal{H}_{s},$ ${\rm Res}_{G,H}(\eta)\neq 0$.

Proof.

Set $a=Ann_{H^{*}(G,k)}(\eta)$.

1. Suppose that ${\rm Res}_{G,H}(\eta)\neq 0$for

some

$H\in \mathcal{H}_{t}$ with$t>0$ andset $b=$ Ann$H^{e}(G,k)({\rm Res}_{G,H}(\eta))$

.

Then

as

${\rm Res}_{G,H}^{-1}(b)\subset a$,

$s=\dim H^{*}(G, k)/a\geqq\dim H^{*}(G, k)/{\rm Res}_{G,H}^{-1}(b)=\dim H^{*}(H, k)/b\geqq$depth $H^{*}(H, k)$

$H$ has

a

central elementary abelian p-subgroup of rank $t$ and depth$H^{*}(H, k)\geqq t$ by

a

theorem of Duflot [5]. Thus we have a contradiction.

2. Suppose that ${\rm Res}_{G,H}(\eta)=0$ for all $H\in \mathcal{H}_{s}$. Then by the statement 1, ${\rm Res}_{G,H}(\eta)=0$

for all $H\in \mathcal{H}_{t}$ with $t\geqq s$. Thus $Ann_{H^{*}(G,k)}(\eta)$ contains $(_{s},$ $\cdots,$$\zeta_{r}$ and

$s=\dim H^{*}(G, k)/$Ann$H^{*}(G,k)(\eta)\leqq\dim H^{*}(G, k)/((S’\cdots, \zeta_{r})\leqq s-1$

This is

a contradiction.

$\square$

2. Results of D.J. Green

In this section we shall prove some results on cohomology algebra of a finite group

having

a

normal elementary abelian p-subgroup. A discussion here depends heavily

on

investigations by D.J.Green [7].

2.1. In this section let $C$ be a finite group with a central elementary abelian p-subgroup

$E$ of order $p^{s}(s>0)$ and

assume

that ap-rank of $C$ is larger than $s$. Set $\overline{C}=C/E$ and

for a subgroup $E\subset H\subset C$, set $\overline{H}=H/E\subset\overline{C}$

.

Set $\mathcal{A}=\mathcal{A}_{1}(E)$ be the set of elementary

abelian p-subgroups $F$ of $C$ containing $E$ with $[F : E]=p$ and $\mathcal{H}=\{C_{C}(F);F\in \mathcal{A}\}$.

And let $\mathcal{K}$ be the family of

subgroups $K\subset C$ such that the Sylow p-subgroups of $C_{C}(K)$

are not conjugate to a subgroup of any of the groups in $\mathcal{H}$ (If $\mathcal{H}$ has a subgroup $H$ with

p’-index in $C$, then weset $\mathcal{K}=\{1\}$

as

before). Notice that if$K\in \mathcal{K}$, then anyelementary

abelian p-subgroup of $K$ is contained in $E$.

Let$\underline{F}_{1},$ $\cdots F_{k}-$, be a complete set of representatives of the C-conjugacy classes in $\mathcal{A}$.

Then $F_{i}\subset C(1\leqq i\leqq)$

are

not $\overline{C}$

-conjugate. let $\sigma_{i}’\in H^{2}(\overline{F_{i}}, GF(p))\subset H^{2}(\overline{F_{i}}, k)$ be a

nonzero

fixed element and set $\sigma_{i}=(\sigma_{i}’)^{p-1}$. Then using the Evens’ norm map, there can

be constructed a homogeneous element $\lambda\in H^{*}(\overline{C}, k)$ such that ${\rm Res}_{\overline{C},\overline{F_{l}}}(\lambda)$ is

a

p-power

of $\sigma_{i}$ for each $i$ (see a discussion in Section 7.1 [7]). Now set $\kappa=Inf_{\overline{C},C}(\lambda)\in H^{*}(C, k)$. $\kappa$ is a primitive element. We know that ${\rm Res}_{C,K}(\kappa)=0$ for any $K\in \mathcal{K}$ and therefore,

replacing $\kappa$ by

some

power, we may

assume

that

$\kappa\in{\rm Im} H_{\mathcal{H},C}$ (see a proofof Lemma 2.6

(3)

Proposition 2.1. The homogeneous element $\kappa\in H^{*}(C, k)$ given above

satisfies

the

fol-lowing properties.

(1) $\kappa\in{\rm Im}$Tr

$H,C$.

(2) ${\rm Res}_{C,H}(\kappa)\in H^{*}(H, k)$ is a regular element

for

every $H\in \mathcal{H}$.

(3) Ann$H^{*}(C,k)(\kappa)=Ker{\rm Res}_{C,H}$

(4) Suppose that $\xi_{1},$

$\cdots,$$\xi_{t}$ is a sequence

of

homogeneous elements

of

$H^{*}(C, k)$ whose

restrictions

form

a regular sequence in $H^{*}(E, k)$. Then (a) $\xi_{1},$

$\cdots,$$\xi_{t}$ is a regular sequence

for

$H^{*}(C, k)/H^{*}(C, k)\cdot\kappa$. In particular, as a

$k[\xi_{1}, \cdots, \xi_{t}]$-module, $H^{*}(C, k)/H^{*}(C, k)\cdot\kappa$ is

free.

(b) KerRes$C,\mathcal{H}$ is

a

free

$k[\xi_{1}, \cdots, \xi_{t}]$-module

(5)

If

the depth

of

$H^{*}(C, k)$ is $s$, then $\kappa$ is

a

zero

divisor in $H^{*}(C, k)$. In particular,

KerRes$C,\mathcal{H}\neq 0$.

Proof.

All the statements inthe proposition except the assertion $(b)$ in thestatement 4

are

included in [7]. We shall prove the stetement 4, $(b)$. Set $R=k[\xi_{1}, \cdots, \xi_{t}]$

.

Then $H^{*}(C, k)$

is

a

free R-module by a result of Duflot [5]. We have the following exact sequence of

R-modules

$0arrow$ Ann$H^{*}(C,k)(\kappa)arrow H^{*}(C, k)arrow\kappa H^{*}(C, k)arrow H^{*}(C, k)/H^{*}(C, k)\kappaarrow 0$

and Ann$H^{*}(C,k)(\kappa)=Ker{\rm Res}_{C,H}$ . Thus by the assertion $(a),$ $4$, it follows that $Ker{\rm Res}_{C,H}$

is a projective R-module. So the result follows because a projective R-module is a free

R-module (see, for example, Theorem 2.5 [8]). $\square$

2.2. In this section let $N$ be a finite group with normal elementary abelian subgroup $E$

of rank $s$ and set $C=C_{N}(E)$. We

use

notations in the previous section for $C$. We shall

prove the following proposition.

Proposition 2.2. Assume that the depth

of

$H^{*}(C, k)$ is $s$. Then

(1) As a $kN/C$-module $Ker{\rm Res}_{C,\mathcal{H}}$ contains a regular module $kN/C$.

(2) $Ker{\rm Res}_{C,\mathcal{H}}$ contains a

nonzero

N-invariant element.

Set $E=\langle a_{1},$ $\cdots,$ $a_{s}\},$ $\alpha_{i}\in H^{1}(E, k)=Hom(E, k)$ be the element dual to

$a_{i}$ and $\beta_{i}=$ $\beta(\alpha_{i})$, where $\beta$ is the Bockstein map. We have a polynomial subalgebra $k[\beta_{1}, \cdots, \beta_{s}]$ in

$H^{*}(E, k)$. Using Evens’ norm map, we obtain homogeneous elements $\xi_{1},$ $\cdots\xi_{s}\in H^{*}(C, k)$

such that ${\rm Res}_{C,E}(\xi_{i})=\beta_{i}^{p^{a}}$ for

some

p-power$p^{a}$. The subalgebra $k[\beta_{1}, \cdots, \beta_{s}]\subset H^{*}(E, k)$

is N-stable, but thesubalgebra $k[\xi_{1}, \cdots, \xi_{s}]\subset H^{*}(C, k)$ may not be N-stable. For$g\in N$,

let $\beta_{i}^{g}=\sum_{j=1}^{s}\lambda_{ij}\beta_{j},$ $\lambda_{ij}\in GF(p)$ and consider the element $\xi=\xi_{i}^{g}-(\sum_{j=1}^{s}\lambda_{ij}\xi_{j})$. Then ${\rm Res}_{C,E}(\xi)=0$ and therefore ${\rm Res}_{C,K}(\xi)$ is nilpotent for any $K\in \mathcal{K}$ because elementary

abelian p-subgroups of such $K$

are

contained in $E$

.

Hence by a theorem of Benson

(The-orem

1.1 [1]$)$

some

p-power of $\xi$ is contained in ${\rm Im} Tr_{H,C}$. Replacing the $\xi_{i}’s$ by suitable

p-powers, we have proved the following.

Lemma 2.3. There exist homogeneous elements$\xi_{i}(1\leqq i\leqq s)$ and ap-power$p^{b}$ satisfying

the following.

(4)

(2) $k[\xi_{1}, \cdots, \xi_{s}]+{\rm Im} Tr_{\mathcal{H},C}$ is N-stable and$k[\xi_{1}, \cdots, \xi_{s}]+{\rm Im} Tr_{\mathcal{H},C}/{\rm Im} Tr_{\mathcal{H},C}\cong k[\beta_{1}^{p^{b}}, \cdots, \beta_{1}^{p^{b}}]$

as

N-modules.

Proof of

Proposition 2.2

Now

we

shall prove the proposition. First

we

prove the statement 1. Set $R_{0}=$

$k[\beta_{1}^{p^{b}}, \cdots , \beta_{s}^{p^{b}}]$ and $R=k[\xi_{1}, \cdots , \xi_{s}]$. Then it is well known that the

sequence

$\xi_{1},$$\cdots,$$\xi_{s}$

is

a

regular sequence for $H^{*}(C, k)$ and $H^{*}(C, k)$ is

a

free R-module (see [5]).

By Proposition 2.1, $Ker{\rm Res}_{C,7\{}\neq 0$. Let $n$ be the least integer such that $V=$

KerRes$C,H\cap H^{n}(C, k)\neq 0$. Then $V$ is N-stable and any k-basis of$V$ canbe extendedto a

set of R-free generators of Ker Res$C,\mathcal{H}$. Then by Lemma 1.2, Proposition 2.1 and the fact

that ${\rm Im} Tr_{7\{,C}$ annihilates Ker Res$C,H$, it follows that $V\cdot R$ is N-stable and $V\cdot R\cong V\otimes_{k}R_{0}$

as

N-modules.

It is known that $R_{0}$ contains $kN/C$

as

a

$kN/C$-module (see [9]).

Therefore

the statement (1) follows.

The statement 2 is

an

easy consequence of the statement 1.

3. Proof of Theorem 0.1

First we shall show that the first statement of the theorem hold. Let $\mathfrak{p}$ be

an

associated

prime in $H^{*}(G, k)$ with $\dim H^{*}(G, k)/\mathfrak{p}=s$ and take

a

homogeneous element $0\neq\eta\in$

$H^{*}(G, k)$ such that Ann$H^{*}(G,k)(\eta)=a$. Then by Lemma 1.2, there exists $E\in \mathcal{A}_{s}$ such

that ${\rm Res}_{G,C_{G}(E)}(\eta)\neq 0$. Set $C=C_{G}(E),$ $\eta_{0}={\rm Res}_{GC)}(\eta)$ and

we

shall

use

the notations in Section 2. Again by Lemma 1.2, $\eta_{0}\in$ KerRes$C,?t$.

Let $\alpha\in H^{*}(C, k)$ such that ${\rm Res}_{C,E}(\alpha)=0$. Then $\alpha\in\sqrt{Ker{\rm Res}_{C\mathcal{K}}}$by

an

argument in

Section

1 and therefore $\alpha\in\sqrt{{\rm Im} Tr_{HC}}$. Thus $\alpha\in\sqrt{Ann_{H^{*}(Ck)}(\eta_{0})}$.

By

a

result of Benson in [1] there exists a homogeneous element $\tau\in H^{*}(C, k)$ such that ${\rm Res}_{N,C}$Tr$C,N(\tau)$ is a regular element in $H^{*}(C, k)$ and restricts to zero on every subgroup

$A$ of $C$ with $A\not\supset E$, where $N=N_{G}(E)$. Set $\sigma=$ Tr$c,c(\tau)$ and consider the element $\alpha=\eta\sigma$. We shall show the following equality hold.

${\rm Res}_{G,C}(\alpha)=\eta_{0}{\rm Res}_{N,C}Tr_{C,N}(\tau)$

As $\alpha=\eta$Tr$c,c(\tau)=Tr_{C,G}({\rm Res}_{G,C}(\eta)\tau)=$ Tr$c,c(\eta_{0}\tau)$, the Mackey double coset formula shows that

${\rm Res}_{G,C}( \alpha)=\sum_{g\in C\backslash G/C}$Tr$c 9\cap c,c{\rm Res}_{C^{g},C^{g}\cap C}((r_{0}\tau)^{g})=\sum_{g\in C\backslash G/C}$Tr

$c9\cap c,c{\rm Res}_{C^{g},C^{g}\cap C}(\eta_{0}\tau^{q})$

Suppose that $C^{-1}\not\supset E$. Then

as

${\rm Res}_{C,C\cap C^{g^{-1}}}(\tau)=0$, it follows that${\rm Res}_{C,C\cap C^{g^{-1}}}(\eta_{0}\tau)=0$

and therefore ${\rm Res}_{C^{g},C^{g}\cap C}(\eta\tau^{g})=0$. If $E\subset C^{g^{-1}}$ and $E\neq E^{g^{-1}}$, then $F=EE^{g}$

‘1

is

elementary abelian, $F\supsetneq E$ and $C\cap C^{g^{-1}}=C_{G}(F)$. So ${\rm Res}_{C,C\cap C^{g^{-1}}}(\eta_{0})=0$ and therefore ${\rm Res}_{C^{g},C9\cap C}(\eta\tau^{g})=0$. Thus

we

have the desired equality. a $\neq 0$ and therefore $\mathfrak{p}=$

Ann$H^{*}(G,k)(\eta)=$ Ann$H^{*}(G,k)(\alpha)$. In these notations, we shall show that for $\rho\in H^{*}(G, k)$, $\rho\alpha=0$ if and only if ${\rm Res}_{G,C}(\rho)\eta_{0}=0$.

Assume that ${\rm Res}_{G,C}(\rho)\eta_{0}=0$, then ${\rm Res}_{G,C}(\rho)\eta_{0}\tau=0$ and $\rho\alpha=$ Tr$c,c({\rm Res}_{G,C}(\rho)\eta_{0}\tau)=$ $0$. Thus we have $\mathfrak{p}={\rm Res}_{G,C}^{-1}($Ann$H^{*}(C,k)(\eta_{0}))$.

(5)

As ${\rm Res}_{\overline{C}^{1}E}(\sqrt{0})\subset\sqrt{Ann_{H^{*}(Ck)}(\eta_{0})}$, we have

${\rm Res}_{G,E}^{-1}(\sqrt{0})\subset{\rm Res}_{G,C}^{-1}($Ann$H^{*}(C,k)(rlo))=\mathfrak{p}$

As ${\rm Res}_{G,E}^{-1}(\sqrt{0})$ is

a

prime ideal and $\dim H^{*}(G, k)/{\rm Res}_{G,E}^{-1}(\sqrt{0})=s$, we can conclude that

${\rm Res}_{G,E}^{-1}(\sqrt{0})=\mathfrak{p}$ and the first statement in the theorem follows.

We next shall prove the second statement in the theorem. Let $E\in \mathcal{A}_{s}$ and

assume

that the depth of $H^{*}(C_{G}(E), k)$ is $s$. Set $C=C_{G}(E),$ $N=N_{G}(E)$ and

we

shall

use

the

notations Section 2. As before let $\tau$ be

a

homogeneous element in $H^{*}(C, k)$ such that

${\rm Res}_{N,C}$Tr$C,N(\tau)$ is

a

regular element in $H^{*}(C, k)$ and restricts to

zero

on

every subgroup

$A$ of $C$ with $A\not\supset E$,

By Proposition 2.2, there exists a non zero homogeneous element $\eta_{0}\in$ Ker Res$C,?t$

which is N-invariant. Notice that ${\rm Res}_{\overline{C}^{1}E}(\sqrt{0})\subset\sqrt{Ann_{H^{*}(Ck)}(\eta_{0})}$. Consider the element $\gamma=\eta_{0}\tau$

.

Then

an

entirely

same

argument

as

above,

we

have

${\rm Res}_{G,C}$ Tr$c,c(\gamma)={\rm Res}_{N,C}$Tr$C,N(\gamma)=\eta_{0}{\rm Res}_{N,C}$Tr$C,N(\tau)$

Set $\alpha=$ Tr$c,c(\gamma)\in H^{*}(G, k)$. We shall show that for $\rho\in H^{*}(G, k),$ $\rho\alpha=0$ if and only if ${\rm Res}_{G,C}(\rho)\eta_{0}=0$.

Assume that $\rho\alpha=0$. Then Tr$c,c({\rm Res}_{G,C}(\rho)\gamma)=0$. Hence ${\rm Res}_{G,C}$Tr$c,c({\rm Res}_{G,C}(\rho)\gamma)=$

$0$. By the similar argument in the above

we

have that

${\rm Res}_{G,C}$ Tr$c,c({\rm Res}_{G,C}(\rho)\gamma)={\rm Res}_{N,C}$Tr$c,N({\rm Res}_{G,C}(\rho)\gamma)=({\rm Res}_{G,C}(\rho)\eta_{0}){\rm Res}_{N,C}$Tr$C,N(\tau)$

As ${\rm Res}_{N,C}$Tr$C,N(\tau)$ is a regular element in $H^{*}(C, k)$, it follows that ${\rm Res}_{G,C}(\rho)\eta_{0}=0$

.

Conversely, if ${\rm Res}_{G,C}(\rho)\eta_{0}=0$, then ${\rm Res}_{G,C}(\rho)\gamma=0$ and $\rho\alpha=$ Tr$C,G({\rm Res}_{G},c(\rho)\gamma)=0$.

A standard argument in commutative noetherian rings says that there exists $\delta\in$

$H^{*}(G, k)$ such that $\eta=\alpha\delta\neq 0$ and $\mathfrak{p}=Ann_{H^{*}(G,k)}(\eta)$ is

a

prime ideal. Then $\eta_{1}=$ ${\rm Res}_{G,C}(\delta)\eta_{0}\neq 0$ and by

an

entirely

same

argument

as

before, we have

Ann$H^{*}(G,k)(\eta)={\rm Res}_{G,C}^{-1}($Ann$H^{*}(C,k)(\eta_{1}))$

As ${\rm Res}_{\overline{C}^{1}E}(\sqrt{0})\subset\sqrt{Ann_{H^{*}(Ck)}(\eta_{0})}\subset\sqrt{Ann_{H^{*}(Ck)}(\eta_{1})},$ ${\rm Res}_{G,E}^{-1}(\sqrt{0})\subset\sqrt{Ann_{H^{*}(Gk)}(\eta)}=$

$\mathfrak{p}$ and $\dim H^{*}(G, k)/\mathfrak{p}=\dim H^{*}(H, k)/$ Ann$H^{*}(C,k)(\eta_{1})=s$. Thus ${\rm Res}_{G,E}^{-1}(\sqrt{0})=\mathfrak{p}$ and

the second statement in the theorem follows.

REFERENCES

[1] D.J.Benson, The image ofthe transfermap, Arch. Math., 61, 7-11, 1993

[2] D.J.Benson, Dickson invariant, regularity and computation in group cohomology, Preprint, 2004

[3] J.F.Carlson, Depth and transfer maps in the cohomology ofgroups, Math. Z., 218, 461-468, 1995

[4] J.F.Carlson, L.Townsley, L.Valero-Elizondo and M.Zhang, Cohomology Rings of Finite Groups,

Kluwer Academic Publishers, 2003

[5] J.Duflot, Depth and equivalriant cohomology, Comment. Math. Helvetici, 56, 627-637, 1981

[6]

[7] D.J.Green, On Carlson’s depth conjecture in group cohomology, Math. Z., 244, 711-723, 2003

[8] H.Matsumura, Commutative Ring Theory, Cambridge Studies in Advanced Mathematics, 8,

Cam-bridge University Press, 1986

[9] P.Symonds, The action ofautomorphisms on the cohomology

of

ap-group, Math. Proc. Camb. Phill.

参照

関連したドキュメント

In [6] we outlined a theory, where certain elements in the Spencer cohomology determine all the complete filtered Lie algebras having a certain graded algebra provided that

arXiv:1101.2075 (2011)), we claimed that both the group algebra of a finite Coxeter group W as well as the Orlik–Solomon algebra of W can be decomposed into a sum of

Thus as a corollary, we get that if D is a finite dimensional division algebra over an algebraic number field K and G = SL 1,D , then the normal subgroup structure of G(K) is given

His idea was to use the existence results for differential inclusions with compact convex values which is the case of the problem (P 2 ) to prove an existence result of the

It is shown that the space of invariant trilinear forms on smooth representations of a semisimple Lie group is finite dimensional if the group is a product of hyperbolic

It is shown that the space of invariant trilinear forms on smooth representations of a semisimple Lie group is finite dimensional if the group is a product of hyperbolic

In addition, we prove a (quasi-compact) base change theorem for rigid etale cohomology and a comparison theorem comparing rigid and algebraic etale cohomology of algebraic

Thus no maximal subgroup of G/P has index co-prime to q and since G/P is supersolvable, this gives, by using a well known result of Huppert, that every maximal subgroup of G/P is